• 现象：exits with return code -7
• 原因：Setting the shm-size to a large number instead of default 64MB when creating docker container solves the problem in my case. It appears that multi-gpu training relies on the shared memory. ref
  • 排查是否是shm-size过小： 在服务器上执行 ds_report，查看最后一行的是不是 shared memory (/dev/shm) size .... 64.00 MB
• 解决方案：增加docker的shm